Finally, given the heterogeneity of the studies, it may not be appropriate to combine them into a summary measure of the effect size.

Anumberofpublishedstudies(21toourknowl-edge)havereportedalterationsinestablisheddiseasegenesfromvarious subgroups of patients with pancreatitis. Unfortu-nately, few investigators have applied the stringent criteria foridiopathic pancreatitis used in our study, and none have in-vestigatedtheextendedfamiliesofpatientswithseeminglyspo-radic trypsinogen mutations. Because this was the specific fo-cus of our study, and because of the limit for references in ourResearchLetter,werestrictedcitationstothemostrelevantworkfor our report.The suggestion by Keim and Teich that our results are dueto inadequate history taking is unfounded since, in contrastwith all prior studies, we obtained both clinical and geneticdata from unaffected family members. This alone allowed usto demonstrate that spontaneous de novo trypsinogen muta-tions must be considered as a cause of sporadic hereditarypancreatitis.Rather than indicating a correlation between frequency oftrypsinogen mutations and impact factors of various journals,we prefer to interpret the Table by Keim and Teich as evi-denceofasignificantvariationinprevalenceofhereditarypan-creatitisamongregionsofdifferentgeneticbackgrounds.There-fore,ourstudycontinuestosuggestthathereditarypancreatitismust be considered as a diagnosis in patients with apparentlyidiopathic pancreatitis—even in the absence of a positive fam-ily history—and specifically in patients younger than 25 yearsat symptom onset. This conclusion is not only supported byother studies